Avoiding the possible data loss error

[daileycon]

What is the best way to avoid the "possible loss of data..." error from this line?

 

int secs_remaining=(Time[0]+PeriodSeconds())-TimeCurrent(); 

[Fernando Carreiro]

daileycon: What is the best way to avoid the "possible loss of data..." error from this line?

int secs_remaining=(Time[0]+PeriodSeconds())-TimeCurrent(); 

Typecasting!

// Either in the "function" style
int secs_remaining = int( Time[0] - TimeCurrent() ) + PeriodSeconds();
// or in the "C" style
int secs_remaining = (int)( Time[0] - TimeCurrent() ) + PeriodSeconds();

[honest_knave]

daileycon:

What is the best way to avoid the "possible loss of data..." error from this line?

 

int secs_remaining=(Time[0]+PeriodSeconds())-TimeCurrent(); 

long secs_remaining=(Time[0]+PeriodSeconds())-TimeCurrent();

[honest_knave]

Fernando Carreiro:
Typecasting!

// Either in the "function" style
int secs_remaining = ( int( Time[0] ) + PeriodSeconds() ) - int( TimeCurrent() );
// or in the "C" style
int secs_remaining = ( (int) Time[0] + PeriodSeconds() ) - (int) TimeCurrent();

Sorry, we posted at the same time. Different ways to skin the proverbial cat!

[Fernando Carreiro]

honest_knave: Sorry, we posted at the same time. Different ways to skin the proverbial cat!

Updated my original post to be a little more efficient!

EDIT: I have to wonder however, if one can have a negative datetime variable, namely is "Time[0] - TimeCurrent()" valid?

EDIT2: Maybe my original post is more correct than my updated "efficient" version:

// Either in the "function" style
int secs_remaining = ( int( Time[0] ) + PeriodSeconds() ) - int( TimeCurrent() );
// or in the "C" style
int secs_remaining = ( (int) Time[0] + PeriodSeconds() ) - (int) TimeCurrent();

[honest_knave]

Fernando Carreiro:

EDIT: I have to wonder however, if one can have a negative datetime variable, namely is "Time[0] - TimeCurrent()" valid?

Yes, a datetime variable can store a negative number (to all intents and purposes, datetime is a long) but if you try to print a negative datetime it will show "wrong datetime"

[Fernando Carreiro]

honest_knave: Yes, a datetime variable can store a negative number (to all intents and purposes, datetime is a long) but if you try to print a negative datetime it will show "wrong datetime"

Yes, but converting a negative "datetime" into an "int" my cause a loss. I've never actually tested that myself!

EDIT: Well, maybe not, because in this case it is just a time difference which perfectly fits an "int".

[honest_knave]

Fernando Carreiro:
Yes, but converting a negative "datetime" into an "int" my cause a loss. I've never actually tested that myself!

In theory, yes. Both ways, positive or negative. Around 68 years / Year 2038 will max out the int.

I just take the hit on the 4 bytes of extra memory and store a long.

[Fernando Carreiro]

honest_knave: In theory, yes. Both ways, positive or negative. Around 68 years / Year 2038 will max out the int. I just take the hit on the 4 bytes of extra memory and store a long.

Hopefully, this "extra" chatter will not confuse the OP!

[honest_knave]

Fernando Carreiro:
Hopefully, this "extra" chatter will not confuse the OP!

Meh, posts 2 and 3 will set them on the right path so they don't have to read very far.

The rest is relevant discussion for those who are interested (probably just you and me!) 

[William Roeder]

honest_knave: The rest is relevant discussion for those who are interested (probably just you and me!) 

This is why I typedef it and make the code self-documenting	#define SECONDS uint #define HR2400 86400 SECONDS    time(datetime when=0){       return SECONDS(when == 0 ? TimeCurrent() : when) % HR2400;               } datetime   date(datetime when=0){       return datetime( (when == 0 ? TimeCurrent() : when) - time(when) );      }